Description Usage Arguments Details Value Note Author(s) See Also Examples
computes polygon segments for TreeCrowns based on watershed algorithm
1 2 3 4 5 6 7 8 9 10 |
chm |
raster - Canopy Height Model RasterLayer. |
a |
numeric - single value for MovingWindow. |
b |
numeric - single value for MovingWindow. |
h |
numeric - maximum height of trees (in meter) to detect trees. |
MIN |
numeric - the minimum area for crowns. Smaller polygons are cropped. |
MAX |
numeric - the maximum area for crowns. Larger polygons are cropped. |
CHMfilter |
numeric - uses a sum filter on the chm with a MovingWindow of (x*x), which must be odd. Default=1 no filter. |
silent |
bolean - if TRUE the function will not print any progress messages (default=FALSE). |
uses a MovingWindow of x*a+b to detect local maxima in a chm to compute TreeCrown Segments
parameter selection - use BestSegVal
to automated detect best fitting parameters for a, b, h, MIN and filter.
returns a PolygonLayer with segments
A 'brute force' segmentation with random parameters is not recommended. TreeSeg is mainly used to compute segments AFTER the validation of best fitting parameters with BestSegVal
.
Further to estimate the quality of the computed polygons it is recommended to use TreeSegCV
for a x-fold CrossValdiation over x different subareas. For full workflow see the 'CENITH_Tutorial'.
Andreas Schönberg
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 | # load packages
require(CENITH)
require(raster)
require(mapview)
# load data
chm <- raster::raster(system.file("extdata","lau_chm.tif",package = "CENITH"))
# take a look on the data
plot(chm)
# NOTE: the example should NOT show to get optimal results (for this see 'BestSegVal')
# start segmentation
x <- TreeSeg(chm,a=0.3,b=0.5,h=1)
length(x)# amount of trees
# compare result with chm
mapview(chm)+x
# clip min and or max polygons
y <-TreeSeg(chm,a=0.2,b=0.5,h=2,MIN=10,MAX=100)
length(y)# amount of trees
mapview(chm)+y
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.